An in-situ spectral calibration method of Thomson ... Meeting... · An in-situ spectral calibration method of Thomson scattering diagnostics for severe radiation circumstances 1H.

An in-situ spectral calibration method of Thomson scattering diagnostics for severe radiation circumstances

1H. Tojo, 2I. Yamada, 2R. Yasuhara, 3A. Ejiri, 1J. Hiratsuka, 3H. Togashi, 3T. Yamaguchi, 1E.Yatsuka, 1T. Hatae, 2H. Funaba, 2H. Hayashi, 3Y. Takase, and 1K. Itami 1Japan Atomic Energy Agency, 801-1, Mukoyama, Naka 311-0193, Japan 2National Institute for Fusion Science, 322-6 Oroshi-cho, Toki 509-5292, Japan 3Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa 277-8561, Japan

1st IAEA Technical Meeting on Fusion Data Processing, Validation and Analysis 1st of June - 3rd of June 2015, Nice, France 1

Data validations and analysis are essential for diagnostic developments

2

accessible ! inaccessible !

lens

Experimental data are correct with enough calibrations.

Experimental data become incorrect.

So far In near future

Use of some special techniques (data acquisition and/or analysis) enables validating data.

This talk focuses on a novel calibration method (data validation) for Thomson scattering diagnostics

Contents

l Thomson scattering diagnostic and significance of calibrations

l Techniques of calibration method l Experimental demonstrations in TST-2 and LHD l Expected performance for JT-60SA, ITER, and

DEMO l Summary

3

Measurements of board spectra are required in Thomson scattering (TS) diagnostic

Board spectrum in high Te

l  Thomson scattering system providing electron temperature (Te) and density (ne) profile

l  Scattered spectra at various Te

4

blue shift

One of fundamental diagnostic for plasma control and study about plasma instabilities

Degradations in optical components have been reported in various tests and experiments

Increase in transmissivity loss

l  Change in fiber transmissivity （neutron irradiation test）

[T. Kakuta et al., J. Nucl. Mater. 1277 307 (2002).]

l  Change due to thin films on the vacuum window surface

[H. Yoshida et al., RSI 68, 256 (1997)]

T=T0exp[-α(λ)t] Decay constant α

5

Transmissivity losses cause underestimation in Te

l Scattered spectra generated from YAG laser

Unknown degraded transmissivity causes underestimations of Te

scattering angle(θ)=125°

Degraded spectrum at 6 keV is close to 5 keV

Developments of calibration methods are necessary for such severe conditions

6

Contents

7

l Thomson scattering diagnostic and significance of calibrations

l Techniques of calibration method l Experimental demonstrations in TST-2 and LHD l Expected performance for JT-60SA, ITER, and

DEMO l Summary

To correct unexpected transmissivity loss, measurements using a standard light source is a good way

8

l Standard light source in the measured position generates known spectrum

Adjustments of sensitivities between the wavelength channels are possible.

However, this method cannot be used under severe radiation condition, requiring some techniques.

[V1, V2, V3, … , V5] polychromator Output signals

transmissivities

I（λ）

known human access is needed!

The known spectrum from wavelength channels are measured.

5 wavelength channels

Principle of a double-pass scattering system

l Measurement using a double-pass scattering system

Scatteings must be measured separately.

Two spectra at the same Te can be evaluated.

l Two scattered spectra from the both scatterings

9

1st pass

2nd pass

1st pass 2nd pass

6080

100120

Out

put [

mV

]

0 50 100 150 200Time [ns]

CH5 (1006-1030 nm)

40

Data analysis: Signal ratios provide correct Te without knowing transmissivities

10

l Measured photon counts at a spectral channel j

the ratio of remains.

Laser energy

Contribution from the spectrum

to be minimized

Measured ratio Theoretical ratio Unknown parameters

Transmissivity(assumed as a constant)

bandpass width scattering length

Te and E2Ls,2/E1Ls,1 are obtained without calibrations.

(No transmissivities !)

When taking ratio of Nm,j between θ = θ1 (first pass) and θ = θ2 (second pass),

Data analysis: Relative transmissivities can be derived using the obtained Te

11

Measured photon counts at j

k=1(first scattering ) 　 2(second scattering)

l Chi-squared to infer relative transmissivities (neηjΔλj)

unknown

11

Cj

can be written by a simple expression

Photon counts from a single pass

showing relative transmissivity between the wavelength channels (j=1,2,3…)

Previous work: Multiple laser method provides Te without knowing transmissivities

l Measurement using two lasers

l Two scattered spectra from YAG and RUBY

Two spectra at the same Te can be measured.

500 600 700 800 900 1000 11000.00.51.01.52.02.53.0

Spe

ctra

l den

sity

Wavelength [nm]

RUBY YAGTe = 5 keV, θ = 140°

[O.R.P. Smith et al., RSI 68, 725 (1996).]

12

Scattered light must be separately measured.

The double-pass scattering system provide overlapped spectra at any Te

13

l Double pass scattering l Multiple (two) lasers

Two spectra can be overlapped at any Te

Two spectra are far from each other. à Te range is limited.

Contents

l Thomson scattering diagnostic and significance of calibrations

l Techniques of calibration method l Experimental demonstrations in TST-2 and LHD l Expected performance for JT-60SA, ITER, and

DEMO l Summary

14

TST-2 and Thomson scattering system

Laser system: YAG laser, 1064 nm, 1.6J Spectrometer: polychromator with 6 wavelength channels

・Plasma current:Ip ~ 100 kA ・Discharge duration: Δt ~ 40 ms ・ ne ~ 1019 m-3 ・Te < 400 eV

Position of Collection mirror

TST-­‐2  

Laser mirror

＜Typical plasma parameters＞ R (major radius): < 0.38 m, a (minor radius):0.25 m, Bt (toroidal field)< 0.2 T,

optical delay

Measured signal ratio provides correct Te

l  Signal ratio between the two passes

SN89151, R = 389 mm

960 980 1000 1020 1040 1060Wavelength [nm]

0

1

2

3

4

5

Ratio

of in

tegr

ated

sign

als

fitted line:Te,c = 196 ± 16 eV

0 50 100 150 200 250Time [ns]

CH5, signal

CH1

CH2

CH3

CH4

CH5CH6

CH4 and CH5 dominate Te determination

Conventional method Te = 205 ± 13 eV

single pass

The fundamental principle of this method was demonstrated.

LHD device and Thomson scattering diagnostic

first pass

second pass

delay optics to measure two scatterings separately

•  Path length ~ 30 m •  Temporal delay ~ 100 ns

LHD experiments already developed a double-pass scattering system to measure high Te and anisotropy in Te

l  Purpose of these experiments Confirm the validity of this method for a higher Te range

I. Yamada et al., Journal of Instrumentation 7 C05007 (2012). K. Narihara et al., Review of Scientific Instruments 72 1122 (2001).

Te range < 2keV

Typical signals from the double-pass scatterings in LHD

l Typical signals Integrated area for Te calculations

No signal in the second pass (forward scattering) à not used

l  Measured ratios and fitting results

Te = 746 eV

CH1

CH4

CH1 CH3

CH4 CH5

Te and the relative transmissivity measurements for a wide Te range have been demonstrated at LHD and TST-2

19

LHD (NIFS)

TST-2 （Univ. of　Tokyo）

Conventional method

Dou

ble-

pass

sig

nal-r

atio

met

hod

l  Te measurements l  Relative transmissivity measurements (LHD)

#116036, t=4500s

measured from standard light source Good agreements between the two

method for a wide Te range　(two orders)

derived

Rel

ativ

e tr

ansm

issi

vity

(std

=CH

3)

H. Tojo et al., 23rd International Toki conference, Toki, Japan, 18th – 21st Nov. 2013

Cj/C

3

submitted to Fusion Engineering and Design.

Contents

l Thomson scattering diagnostic and significance of calibrations

l Techniques of calibration method l Experimental demonstrations in TST-2 and LHD l Expected performance for JT-60SA, ITER, and

DEMO l Summary

20

JT-60SA device

21

Toroidal Field Coils

Cryostat

Divertor

Central Solenoid Coils

Equilibrium Field Coils

Vacuum Vessel

NBI

Plasma Current Ip 5.5MA Toroidal Field Bt 2.25T Major Radius Rp 2.97m Minor Radius ap 1.18m Elongation κx 1.93 Triangularity δx 0.5 Safety Factor q95 3 Plasma Volume Vp 133m3 Flat Top 100 s

Total heating power (NB+ECH): 41MW

JT-60SA utilizes super conducting toroidal / poloidal coils, new vacuum vessel and new cryostat.

The double-pass scattering diagnostic will be utilized in JT-60SA

22

Target spec. core edge

Radial spatial resolution

2 – 3 cm ~ 1 cm

Dynamic range 0.1 – 30 keV 0.01 – 10 keV

l  port-P2 collection optics (core)

l  port-P5 collection optics (high field side, edge)

Equatorial plane of JT-60SA

l  ❑ port-P1 collection optics (low field side, edge)

Plasma

H. Tojo et al., RSI 81 10D539 (2010).

YAG laser

Spatial channel at the core (θ = 125°) was used to evaluate accuracy

Systematic errors in Te can be negligible with this new method

23

l  Procedure to evaluate calibration accuracy

Nm,1 = Ns,1 + Random No. (-σj,k – +σj,k )

Nm,2 = Ns,2 + Random No. (-σj,k – +σj,k )

Calculation of Te and Cj for 1000 times

Ideal scattered photon counts

Fiber w/o irradiation　(standard)

irradiated fiber 　(measured in JT-60U)

l  Assumption: unknown degradation occurs

unknown change

50

100

150200

Cou

nts

TRUE:10.00keV

GaussianCenter: 8.76keVSig (%): 2.08

8.0 8.5 9.0 9.5 10.0 10.5Electron temperature [keV]

050

100

150C

ount

sGaussianCenter: 10.00keVSig (%): 2.82

l  Histograms of calculated

Conventional method

Double-pass (Signal ratio)

Data validity is confirmed.

Relative transmissivity losses can be determined with enough accuracy

Rel

ativ

e er

ror i

n C

j

1062.5–1065.5nm

Low signal intensity

500 – 660 nm

C4/1019

Cou

nts

Wavelength channel

l Distribution of relative transmissivity

Error

In JT-60SA, we are now planning to use this method for not only correct Te determination but also monitoring trasmissivity change.

Use of one port is a realistic design of Thomson scattering diagnostics in ITER and DEMO

25

l Geometry of TS system using one port

Laser and collection optics are closed each other.

Calibrations using a double-pass scattering system in such wide scattering angle and high Te have not been considered.

few spaces for diagnostics

One port !!

Reflection mirror is necessary but need various analysis. (EMF, heat load, and oscillations)

high Te plasma

Optimization of wavelength channels are essential to measure both spectra.

400 600 800 1000Wavelength [nm]

0

1

2

3

4

Spec

tral d

ensi

ty

Forward (θ = 180°-160° = 40°)

Backward (θ = 160°)

Te = 40 keV

•  The number of wavelength channels •  Measured wavelength range

26

(1) Optimize wavelength channels considering both scatterings

to find the best configuration for good accuracy

(2) Estimate Te and Cj accuracies using the optimized conditions

l Scattered spectra at wide scattering angle (θ1=160° and θ2=40°)

l Accuracy evaluations are performed by the following two steps

Configuration of wavelength channels must be carefully selected.

Calibration methods of ITER Thomson scattering system and parameters used in this study

27

l Core measurements l Edge measurements

Calibration method G.S. Kursliev et al., Nuclear Fusion 55 053024 (2015).

E. Yatsuka et al., J. Plasma Fusion Res. SERIES 9 12 (2010).

Two or three lasers Two lasers

Accuracy

Scattering angle

Te range

θ < 160° θ < 140°

< 10 keV < 40 keV

Te cannot be evaluated in some Te range (2 lasers)

Low Te cannot be evaluated.

ß In this study, parameters for the core measurements are used to evaluate accuracies.

Scattering length (Ls) 63 mm

Scattering angle (θ) 160° Solid angle (Ωs) 2.72 msr Transmissivity except for polychromator (Topt)

35%

Target normalized radius ((R-R0)/a)

-0.1

Use of border enable finding good segments of wavelength channels

Err.(λ1,λ2,λ3) =σ Te

Te

Segment

averaged error over the Te range

# of channels (Nch): 5

The minimizations were performed by the simulated annealing method [2]. W. Press et al., Numerical Recipes in Fortran, second edition, Cambridge University Press (1992).

l Locations of the wavelength channels

l  Error calculation

Optimizing to find the minimum error

28

O. Naito et al., Rev. Sci. Instrum. 70 (1999) 3780.

0.01 0.10 1.00 10.00 100.00Te [keV]

0.01

0.10

Aver

age

rela

tive

erro

r

Optimized wavelength channels provide good accuracy (<10%) for Te = 0.5 – 40 keV , same level as three laser method

[λ1, λ2, λ3, λ4, λ5] = [921, 995, 1034, 1049, 1057 nm]

σ Te

Te= 5.3%

l Averaged relative errors in Te at Nch = 7

θ = 160°

l Scan over the No. of wavelength channels

Minimum error at Nch =7

(Nch)

No clear change for Nch > 5 is attributed to internal transmissivity losses in polychromators

3 4 5 6 7 8# of wavelgnth channels

0.000.050.100.150.20

Aver

age

rela

tive

erro

r R

elat

ive

erro

r in

T e

G.S. Kursliev et al., Nuclear Fusion 55 053024 (2015).

l  ITER core TS

Two lasers

Three lasers

Rel

ativ

e er

ror i

n T e

0.1 1.0 10

0.1

1

Te [keV]

Relative transmissivities (calibration factors) can be measured

All wavelength channels can be measured in Te = 1 keV measurements. àpossible to monitor degradation even during experiments

30

l  Cj with expected error bars at Te= 1keV

l  Relative errors of Cj at Te= 0.1, 1, 40 keV

Contents

l Thomson scattering diagnostic and significance of calibrations

l Techniques to overcome radiation conditions l Experimental demonstrations in TST-2 and LHD

l Applicability in JT-60SA l Expected performance for ITER and DEMO l Summary

31

Summary

l  An in-situ calibration method using a double-pass scattering system has been developed as a data validation technique. •  Measurements of two overlapped spectra enable providing Te without

knowing the transmissivities. l  Te and relative transmissivities have been successfully measured in

TST-2 and LHD (0.1 – 2 keV). l  JT-60SA will utilize a double-pass scattering system

•  The data validation is confirmed by numerical analysis showing the method suppresses a systematic error in Te.

l  This method can be used in new type of Thomson scattering

diagnostic with a wide scattering angle and high Te. •  Wavelength channels considering the forward and backward scattering

are optimized to provide high accuracy in Te. •  Good accuracies (<10%) for a wide Te range are expected.

32